SLAtech General-Business
71/100Baseline reference; vertical-routing native, RTL Hebrew polish
General-business eval scoreboard
SLAtech General-Business: 71/100
Reproducible 200-question general-business eval harness. Baseline reference против которой все per-vertical specialised models compared. Пара с umbrella eval scoreboard, general-business glossary и general-business FAQ.
Score breakdown по category
| Category | General-Business | Generic competitor | Gap |
|---|---|---|---|
| Lead-capture quality Structured name + company + role + use-case intake. Generic chatbots collect contact form fields only — нет qualification depth. |
78 | 68 | +10 |
| General FAQ coverage Hours / pricing / policy queries hit cached answer cards. Generic chatbots match здесь когда given clean FAQ source. |
76 | 71 | +5 |
| Routing к specialist vertical Self-classification: identifies industry-vertical из query и routes к relevant Med/Edu/Legal hub. Generic chatbots не имеют vertical-routing intelligence. |
80 | 52 | +28 |
| Multilingual support (HE / RU) Generic chatbots score slightly higher здесь due к broader auto-translate coverage — General-Business does not invest в vertical-specific HE/RU terminology. |
68 | 70 | -2 |
| Confidentiality / PII posture Basic PII redaction at ingest. Vertical models (Med, Legal) ship deeper redaction + BAA/UPL workflows. |
65 | 60 | +5 |
Competitor comparison
Intercom Fin (generic)
64/100Mid-market polish но English-first, нет vertical-routing, нет eval published
Drift (generic)
67/100Strong demo-booking но English-first, US-locale defaults
Tidio Lyro (generic SMB)
58/100SMB-priced но нет Hebrew RTL polish, conversation cap на lower tiers
Продолжите оценку покупателя
Per-vertical eval score — один input. Три других инструмента самообслуживания закрывают картину без звонка с продавцами:
Reproduce eval против вашего tenant
Eval methodology — open-source. 200 sealed general-business questions с LLM-as-Judge scoring на factuality, hallucination и confidence axes.